**** BEGIN LOGGING AT Wed Sep 21 02:59:57 2011 Sep 21 16:43:09 sgw? Sep 21 16:44:33 fray, good morning Sep 21 16:44:57 sudo is broken when pam is enbled -- it doesn't have a dependecy on PAM.. Sep 21 16:45:07 are we doing any core-image-lsb builds in the autobuilder? Sep 21 16:46:23 fray, yes and the CDC team is supposed to be testing them. Sep 21 16:46:39 Hmm.. wonder why I found the issue then Sep 21 16:46:58 "nothing provides pam"? huh Sep 21 16:47:18 ahh libpam Sep 21 16:48:31 fray: i do lsb builds all the time Sep 21 16:50:11 I suspect I have a greater number for my parallel build or something.. polkit seems to be missing the dep as well.. Sep 21 16:50:24 so just sudo and polkit are issues Sep 21 16:50:42 nope, I'm wrong.. it does have libpam in the depends.. Sep 21 16:50:45 so only sudo was missing it Sep 21 16:52:53 looks like hte patch should always be enabled... Hmm.. (anyone know anything about polkit?) ;) Sep 21 16:53:51 fray: not directly, but your right, sudo seems to be missing libpam, does it need to be conditional on DISTRO_FEATURES? Sep 21 16:54:56 ya.. Sep 21 16:55:09 +DEPENDS += " ${@base_contains('DISTRO_FEATURES', 'pam', 'libpam', '', d)}" Sep 21 16:55:09 + Sep 21 16:55:17 I'll submit a patch in a bit Sep 21 16:55:41 fray: which package is failing? Sep 21 16:56:15 fray: i hit some similiar sounding issues with 'at' and that DISTRO_FEATURES/pam stuff Sep 21 16:56:31 it was sudo Sep 21 16:56:49 at appears to be implemented right from my quick glance Sep 21 16:57:26 it was not liking something at the time Sep 21 16:57:29 sgw, do you thing sudo needs specific pam RDEPENDS? Sep 21 16:57:33 Hmm.. Sep 21 16:57:59 I see other pam enabled packages with them Sep 21 16:58:36 fray honestly I do not know Sep 21 16:59:47 I'll do some basic investigation in a bit and verify the libpam depends are reasonable for these items Sep 21 17:00:05 I suspect everything else is reasonable and sudo needs an RDEPENDS Sep 21 17:00:19 (specifically to require the pam module it needs/is configured to use) Sep 21 17:00:39 It looks like in some cases the RDEPENDS_${PN} is a duplication of the DEPENDS which is really un-needed (polkit example) Sep 21 17:00:48 yes, I agree.. Sep 21 17:01:03 if the thing uses libpam, well it's going to automatically pickup libpam from the soname processing Sep 21 17:03:19 fray: but the depends still needs to be there I think for configure to process it correctly, the RDEPENDS in polkit and at specifically look redundant Sep 21 17:04:55 yes.. DEPENDS is needed.. it's RDEPENDS that shouldn't be needed Sep 21 17:05:11 unless we depend on something that can't be satisfied via the automatic system (such as pam plugins) Sep 21 17:05:23 we likely could right a dependency generation for the pam configuration files.. ;) but not today Sep 21 17:14:16 ohh something else blew up... Sep 21 17:23:58 Hmm foomatic-filters failed Sep 21 17:24:27 checking for texttops... no Sep 21 17:24:28 file-converter: Sep 21 17:24:28 configure: error: cannot find a2ps, enscript, mpage, or CUPS' texttops. You need to have at least one installed Sep 21 17:24:54 Hmm.. wonder if that is a host contamination issue Sep 21 17:25:37 fray: what about this change looks like: http://git.yoctoproject.org/cgit/cgit.cgi/poky-contrib/commit/?h=dxu4/ml-testing Sep 21 17:25:46 looking Sep 21 17:26:40 yes, that looks good Sep 21 17:26:59 ok, I will send out a pull request for that. thanks:) Sep 21 17:27:10 (I assume it's working better at least?) Sep 21 17:28:24 fray: compared with the last version, what we change is mainly removing the line of "rm -rf xxx/install" Sep 21 17:28:38 yup.. and that should be fine.. Sep 21 17:29:05 check that the rootfs doesn't contain "/install".. and that work/.../temp/ contains each fo the solution files for debugging.. (where ... is the image tyoe) Sep 21 17:29:33 fray: yes, it behaves as you described Sep 21 17:29:47 excellent! Sep 21 17:32:22 sgw: I am trying to reproduce the sato-sdk rootfs error in my local machine Sep 21 17:34:58 dongxiao: OK Sep 21 17:38:28 I need to set some environment vairables before configure is called.. what is the right way to do that? do_configure_prepend? Sep 21 17:40:38 fray: yes I think so. refer to eglibc-initial_2.13.bb, there is a do_configure_prepend() function which unset an environment variable. Sep 21 17:40:59 thanx Sep 21 17:50:13 hello. I just got install done and working for armqemu. Great!! Sep 21 17:50:33 now I want to do the beagle board. Iam a newbie Sep 21 17:50:44 sgw: Hi Sep 21 17:50:48 can I edit local.conf and switch to beagle without recreate the project or build Sep 21 17:50:48 everything from scratch Sep 21 17:51:00 sgw: please take the three commits I did to X inc files Sep 21 17:51:12 sgw: those are trivial and reduce my delta ;-) Sep 21 17:51:14 alyneho, due to the tune files for the beagleboard (cortex a8) you are going to be rebuild most things Sep 21 17:51:44 hmmm ... ok better to just create new project? Sep 21 17:53:39 alyneho, new project? Sep 21 17:54:02 alyneho, you mean a new build directory? there is no real need for that, you can just change MACHINE=beagleboard and build Sep 21 17:54:37 Ok .. I will do that. Thanks!! Sep 21 17:56:15 I need a clue.. do_configure_prepend and EXTRA_OECONF don't seem to be working for me.. Sep 21 18:23:48 quit Sep 21 18:43:38 RP__: http://patches.openembedded.org/bundle/otavio/xf86-fixes/ Sep 21 18:44:07 RP__: I prefer to not send it again since it is waiting for apply for more then a week Sep 21 18:44:17 RP__: those are trivial enough to go in, I think. Sep 21 18:55:06 how to start poky-minimal image to boot.... I 've both minimal and sato images. When I start it is taking sato image by default Sep 21 18:58:01 using qemu? Sep 21 18:58:12 specify the rootfs you want.. i.e. Sep 21 18:58:24 runqemu qemux86 core-image-minimal Sep 21 19:07:37 fray, thanku! Sep 21 19:07:55 np Sep 21 19:09:15 $ poky-qemu Sep 21 19:09:32 I used that. In my system runqemu is not working!! Sep 21 19:10:20 It is sayig command not found!! Sep 21 19:10:21 whats the failure? Sep 21 19:10:55 Are you using master or a released version of Poky? Sep 21 19:11:09 sgw: I found a few more failures on the LSB build Sep 21 19:11:49 usbutils wants pkgconfig (runtime).. python-netserver wants "/usr/local/bin/python" Sep 21 19:12:04 perl-misc wants the sysroot perl-native version.. :( Sep 21 19:12:11 and ncurses-dev was dragged in Sep 21 19:12:44 FILES_${PN} += "${datadir}/pkgconfig" Sep 21 19:12:46 bingo Sep 21 19:13:42 I 'm using along with yocto Sep 21 19:14:05 Arial700 there was a rename a while back.. I'm trying to understand ifyou are using the newer version or now.. Sep 21 19:14:26 when I run it, I run /runqemu Sep 21 19:14:48 so in my build directory... ../scripts/runqemu qemux86_64 core-image-minimal nographic Sep 21 19:15:11 if it tells me it can't get to qemu, I simply do a bitbake qemu-native and that corrects it.. Sep 21 19:15:33 ohk.. I tried to run it as if like a shull command like $ runqemu Sep 21 19:16:03 I don't know if runqemu is in the path or now.. I always run it directly myself Sep 21 19:16:14 (habbit if for no other reason) Sep 21 19:17:04 fray, gr8! but can u tel me why do we run bitbake qemu-native ? what does it do? Sep 21 19:17:20 that will build the native (host version) of qemu Sep 21 19:17:36 if for some reason qemu wasn't built as part of the image creation, you may have to build it manually for the runqemu step Sep 21 19:17:56 fray, ohk.. got it ... Thanku Sep 21 19:18:19 np Sep 21 19:18:25 can u tel how can i transfer data between host system and qemu image Sep 21 19:18:51 If I install samba.. it is going to be heavy Sep 21 19:18:52 if the network interface is supported on your system, you can scp/ssh it.. you'll have to look at ifconfig (or equivalent) for the right IP address.. Sep 21 19:19:23 ! So I want to 've to go it in a light weight mode.. :) Sep 21 19:19:28 alternatively you can setup the NFS server.. and the QEMU target will use a directory on your host side for booting.. (I havn't set this up in a while, so I don't remember how to do that) Sep 21 19:19:30 Ya.. I 'll try ssh way Sep 21 19:19:47 but when you do setup that you can simply copying and move files on the host side, and they'll appear on the target for development Sep 21 19:20:08 I believe the Yocto instructions explains the NFS server setup.. (You don't need to be root for the NFS serving BTW..) Sep 21 19:22:07 fray, gr8! Sep 21 19:24:45 Can we get the source code for the sato project Sep 21 19:25:31 I need the applications given in that project! Gnome website does not 've any single word about Sato! Sep 21 19:28:46 the sources are in the downloads directory Sep 21 19:29:07 fray, Ohk.. I'll check now Sep 21 19:29:30 does anyone know what I have to clean to get bitbake to include newly added packages to MACHINE_ESSENTIAL_RDEPENDS in the image it builds? Sep 21 19:37:32 fray, I could not find the sources!! there were only some svn files (desktop entires) and icon themes about sato Sep 21 19:38:09 downloads/*sato* Sep 21 19:41:52 desktop-folders_svn.o-hand.com_.repos.sato.trunk_76_.tar.gz Sep 21 19:41:52 desktop-folders_svn.o-hand.com_.repos.sato.trunk_76_.tar.gz.done Sep 21 19:41:52 gtk-engine_svn.o-hand.com_.repos.sato.trunk_163_.tar.gz Sep 21 19:41:52 gtk-engine_svn.o-hand.com_.repos.sato.trunk_163_.tar.gz.done Sep 21 19:41:52 matchbox-theme-sato-0.1.tar.gz Sep 21 19:41:53 matchbox-theme-sato-0.1.tar.gz.done Sep 21 19:41:54 poky-image-sato.bb.done Sep 21 19:41:58 sato-icon-theme-0.4.1 Sep 21 19:42:01 sato-icon-theme-0.4.1.tar.gz Sep 21 19:42:02 sato-icon-theme-0.4.1.tar.gz.done Sep 21 19:42:04 task-poky-x11-sato.bb.done Sep 21 19:42:11 Those were the files which has sato name.. !! Sep 21 19:45:38 yes.. the .gz are the archives with the sources Sep 21 21:08:46 fray: are you expecting to see some rpm dep errors on latest git head? Sep 21 21:09:01 | Processing task-base-extended... Sep 21 21:09:01 | error: Failed dependencies: Sep 21 21:09:01 | /local/home/mattsm/git/poky/build_p4080ds_release/tmp/sysroots/x86_64-linux/usr/bin/perl-native/perl is needed by perl-misc-5.12.3-r4.ppce500mc Sep 21 21:09:01 | /usr/local/bin/python is needed by python-netserver-2.6.6-r2.9.ppce500mc Sep 21 21:09:04 depends on the img type Sep 21 21:09:09 lsb Sep 21 21:09:14 and yes.. I have a fix for those n my local tree Sep 21 21:09:17 k Sep 21 21:09:36 still working few a few remaining items though Sep 21 21:10:08 it's likely going to be 2-3 more hours here before I have anything useful though.. Sep 21 21:10:17 I'm trying to figure out what is dragging ncurses-dev into the image Sep 21 21:10:22 (and other -dev packages Sep 21 21:10:39 no big rush here, just wanted to ask Sep 21 21:21:13 are you supposed to build pseudo-native first or should it do that automatically then run the target you asked for to begin with? Sep 21 21:31:55 msm, if you use the bitbake "script" it will do that for you.. if you don't use the script.. you need to start -- otherwise pseudo may not be running when it should be.. ending up with packages with incorrect owners/groups/etc.. Sep 21 21:32:21 scripts/bitbake is the script.. should be in your path Sep 21 21:32:35 im using that Sep 21 21:32:41 it just builds pseudo-native then stops Sep 21 21:32:49 then yes.. it will build pseudo-native. or it should Sep 21 21:32:49 if i rerun the same command w/ the wrapper it works Sep 21 21:33:06 but it should not just stop i suspect, it used to keep building the actual target you asked for Sep 21 21:33:11 I don't know.. it builds it for you and then continues on with the command you executed.. (thats what it's suppoosed to do) Sep 21 21:33:36 you'll likely need to figure out why it's stopping.. the script is simple enough, just run it with bash -x and you should see where it stopped Sep 21 21:33:45 yea Sep 21 21:39:13 its the return value from bitbake proper Sep 21 21:39:34 from Sep 21 21:40:24 bitbake pseudo-native $TARTARGET -c populate_sysroot Sep 21 21:40:24 ret=$? Sep 21 21:40:24 if [ "$ret" != "0" ]; then Sep 21 21:40:24 exit 1 Sep 21 21:40:24 fi Sep 21 21:40:46 i've noticed bitbake return codes are weird lately Sep 21 21:40:58 if thats the case, it's possible soeone broke bitbake... :( Sep 21 21:41:03 well that they are not always 0 when the thing finishes Sep 21 21:41:19 well they should be non-zero on failure...... Sep 21 21:41:29 if pseudo-native didn't fail though.. 0 is what I would expect Sep 21 21:41:40 pseudo-native passed Sep 21 21:43:10 not sure.. I've got to run.. but I think I have a patch set to fix LSB building.. I'll clean it up and push it to my contrib tree when I get back Sep 21 21:49:36 whats the proper place to document this? bitbake bugzilla or…? Sep 21 22:00:38 nvm, i think i have a small fix **** ENDING LOGGING AT Thu Sep 22 02:59:56 2011